goldenarm

It's impressive to see the battery capacity growing on ElectricityMaps over the years : https://app.electricitymaps.com/map/zone/US-CAL-CISO/yearly

But why is the utilization rate so low? It was barely 0.8% of electricity in California last year.

Edit: Got the answer, the batteries fire up around 6pm during the evening peak at around 9GW for ~25% of the total consumption. We would need approximately 5x more batteries and renewable power to move out of carbon-intensive production entirely.

Since the robocall tsunami of the 2020s, people stopped answering phone calls entirely.

The same thing might happen to the entire internet.

The "massive technological improvements" already happened.

Since the 1960s, the worker-to-retiree ratio halved indeed, but productivity went 4x.

Impressive work. I wish they did it on Qwen3.5 though, because their beefed up Qwen3 8B is still way worse than Qwen3.5 9B.

Has Google (or other adtech) ever been sued for this?

They do barely any due diligence and are basically distributing malware and badware at a massive scale without any consequences.

Google shutting down it's public cache made it even worse.

"I believe that AI could [...] usher in a renaissance of democracy and freedom"

How exactly? So far AI has accelerated misinformation at scale and wealth concentration.

Between this and Huggingface, when will any victim sue OpenAI for this ?

Original Source : https://mathstodon.xyz/@tao/117237320796901560

I recommend Terrence Tao's commentary on such a proof : https://mathstodon.xyz/@tao/117219101339291693

Key quote : "Solving the problem by purely AI-powered methods [would be a] net negative for the progress of mathematics."

Key quote : "Solving the problem by purely AI-powered methods [would be a] net negative for the progress of mathematics."

@dang please can we add a [2014] to the title ?

Thank you, your comment was more informative than the economist article

The article dismisses everything without sourcing anything.

US electricity is 40% more expensive since 2022 https://fred.stlouisfed.org/series/APU000072610

Some AI datacenters can indeed be heard a mile away https://www.theguardian.com/us-news/2026/aug/28/datacenters-...

Amazon's new Texas AI data center could become the biggest CO2 polluter in the US https://www.techradar.com/pro/amazons-new-texas-ai-data-cent...

The only benchmark I don't want to be saturated : https://felonybench.com/

I love Gemini but god their SEO is awful, it's very difficult to look for anything related to it. They should consider a new name

Reminds me of the "entreprise grade fizzbuzz"

https://github.com/enterprisequalitycoding/fizzbuzzenterpris...

You're describing the 2010s era of algorithmic social media, I think OP is referring to the earlier internet

Many are on old laptops, which suffer the same way

Astra is >10TB and might struggle to self replicate, but the wicked-smart qwen3.8 27B is 20GB and could easily spread on botnets

I'm confused by the situation. I'm the last manual coder of my company, and am shipping projects faster than my colleagues who are spending fortunes in tokens.

I was intrigued by the hype and gave a chance this week to codex+sol 5.6 and cc+opus 5. They cheated, lied, disobeyed, and shipped subtle bugs so often, it wasted more time that if I did it myself.

Is half of the industry under AI psychosis right now ? Will models get better ?

My relatives and friends are fed up with enshittification and hostile tech, some are about to stop using tech entirely.

Big tech is too short-termist to realize it, but we are about to destroy a lot of economical value if we destroy the internet and the trust behind it.

Geometric mean of all these benchmarks :

* GPT-5.6 Sol: 65.5

* Fable 5 (w/ fallback): 64.5

* Opus 5: 64.0

* DS-V4-Pro 0813: 62.5

* Kimi-K3: 62.3

* DS-V4-Flash 0731: 55.8

* GLM-5.2: 47.3

I respect Anthropic for dogfooding and vibecoding their own products.

The unfortunate consequence is low quality engineering and a billion dollar product with 15k pending Github issues.

PS5 prices went up 35% in the past 2 years, and PC RAM went 5x. This is a global problem

They killed greader and RSS to push Google+, which was a brilliant decision.

Why are models better than agents, isn't it supposed to be the opposite? I don't understand the difference and what you are measuring.

Turtles with law degrees don't exist, but LLMs that increase tech debt in code bases do. Maybe recent ones got better, and it's useful to quantify the progress being made.

It's not an invention it's a question. If the number is <20% that's great news for the Chrome team.